Usefulness of a Carbon target in DUNE ND : first thoughts DUNE ND - PowerPoint PPT Presentation

Usefulness of a Carbon target in DUNE ND : first thoughts DUNE ND meeting – 15 May 2019 S.Bolognesi – IRFU/CEA

Why neutrino-nucleus interactions are important Modeling of neutrino-nucleus interactions is needed for ND → FD extrapolation because of: - different E n energy spectrum at ND (before oscillation) and at FD (after oscillation) ND constrain only the convolution of xsec and flux → need to disentangle them to extrapolate correctly from ND energy to the oscillated energy spectrum at FD - reconstruction of neutrino energy from particles observed in the final state - extrapolation between different neutrino species: mostly n m , n m at ND → need also n e and n e at FD need to measure xsec and flux of n m and n m at ND to minimize model-dependence - extrapolation between different acceptances at ND and FD (due to different size) - extrapolation between different nuclear target: even for same active target in the fiducial volume at ND and FD, still different composition for background coming from out-of-fiducial volume usage of different nuclei is a handle for better understanding of neutrino-nucleus interactions ! 2

Dependence on E n FD ( E ν )σ ν α ' ( E ν ) dE ν ND ≈ ∫ oscillated flux P ν α → ν α ' ( E ν )ϕ ν α ' FD N ν α ' N ν α ND ( E ν )σ ν α ( E ν ) dE ν ∫ unoscillated flux ϕ ν α ' What we actually constrain is the probability of a given final state observed in the detector e.g. Rate ( m +p+ p + ) is actually the convolution of: cross-section of probability of finding probability for the fundamental EWK X the proton in the proton/pion to exit X interaction nucleus (and extract it) the nucleus s ( n +p → m +p+ p + ) Different component of the cross-section: intial state nuclear effects (IS), fundamental EWK interaction ( s ), final state interaction (FSI) in the nucleus. Each component has a different neutrino energy dependence E.g.: a final state without pion can be due to a CCQE event or to CCRes pion production followed by FSI absorption of the pion → if FSI is wrongly estimated, the extrapolation to the far detector is wrong because the energy dependence of CCQE and CCRes xsec is different This is a particularly complex problem in a wide-band beam where many different processes (CCQE, 2p2h, CCRES, Multipion, DIS) have all large xsec Need to separate each component IS/ s /FSI in order to extrapolate them correctly from ND measurements to far detector 3

Reconstruction of E n FD ( E ν )σ ν α ' ( E ν ) F theo ( E vis − E ν ) dE ν FD ( E vis ) ≈ ∫ oscillated flux P ν α →ν α ' ( E ν )ϕ ν α ' N ν α ' ND ( E vis ) N ν α ND ( E ν )σ ν α ' ( E ν ) F theo ( E vis − E ν ) dE ν ∫ unoscillated flux ϕ ν α ' We need to go from the observed particles in the final state to the incoming neutrino energy Again we need to control each component separately IS/ s /FSI to get it right: ● initial state effects: e.g. energy lost in the nucleus (“binding energy”) ● fundamental interaction: e.g. CCQE (p final state) vs 2p2h with neutron component (pn final state) ● final state effects: e.g. proton deceleration, pion absorption... We need to correct for each of these effects (IS/ s /FSI)! 4

How to constrain IS/ s /FSI New kind of observables including the proton (neutron) information I will use single transverse variables as a proxy: many more can be thought (p n , Ehad, vertex activity...) I will mostly discuss protons, neutrons, similar arguments holds for pions arXiv:1901.03750 ● The bulk of d p T is sensitive to initial state effects: Fermi momentum distribution ● Fundamental interaction: separate CCQE from 2p2h d p T tail ● What about FSI ? 5

How to constrain IS/ s /FSI New kind of observables including the proton (neutron) information I will use single transverse variables as a proxy: many more can be thought (pn, Ehad, vertex activity...) I will mostly discuss protons, neutrons, similar arguments holds for pions arXiv:1901.03750 da T is sensitive to FSI : how much acceleration/deceleration of the proton in the nucleus → da T shape (~flat without FSI) 6

Usefulness of Carbon The capability of separating the different effects (IS/ s /FSI) in these variable is only 'partial', there is always some degeneracy in the shapes between the different effects Measurement of da T / d p T for different targets help disentangling IS/ s /FSI effects! Since they all have a different dependence on nucleus size A Difference between C vs Ar give enough leverage for extracting A-depending effects separately ➢ FSI can be extracted from da T shape: preliminary parametrization of A-dependence can be extracted from electron scattering data and further tuned with ND data 7

Usefulness of Carbon The capability of separating the different effects (IS/ s /FSI) in these variable is only 'partial', there is always some degeneracy in the shapes between the different effects Measurement of da T / d p T for different targets help disentangling IS/ s /FSI effects! Since they all have a different dependence on nucleus size A Difference between C vs Ar give enough leverage for extracting A-depending effects separately ➢ Initial state effects (Fermi momentum) can be extracted from the width of the d p T distribution (other variables are sensitive to binding energy) Fermi momentum dependence on A from electron scattering SuSaV2 model: these values applied to Relativistic Mean Field model assure scaling of 2 nd kind in the super-scaling 8 functions for neutrino scattering Phys. Rev. C 71, 065501

Usefulness of Carbon The capability of separating the different effects (IS/ s /FSI) in these variable is only 'partial', there is always some degeneracy in the shapes between the different effects Measurement of da T / d p T for different targets help disentangling IS/ s /FSI effects! Since they all have a different dependence on nucleus size A Difference between C vs Ar give enough leverage for extracting A-depending effects separately ➢ Fundamental interaction, eg. 2p2h/CCQE, affect the height of peak/tail in d p T 2p2h and CCQE cross-section have different A dependence (e.g. SuSa model: 2p2h ~ A*k F 2 , CCQE ~ A/k F ) A-dependence of the cross-section is a powerful handle to evaluate CCQE and 2p2h separately (thus extrapolating properly the xsec from ND to FD) 9

Usefulness of Carbon: further steps Need to be more quantitative: ● Quantify IS vs FSI precision with multidimensional fit in 3DST-like detector (spoiler: % level accuracy!!) ● Study how to combine C and Ar target: is our 'cascade' semi-classical model enough? → interesting existing electron scattering data to explore 10

E n reconstruction: neutrons (1)  Big advantage of DUNE is the capability of reconstructing the total final state energy as a proxy of the incoming neutrino energy While this minimize the model-dependence of the E n reconstruction, same model- dependence still remain (binding energy, neutrons ...)  The modeling of the 'hadronic' part of the final state (all what is not the lepton) is (almost) terra-incognita First 'calorimetric' measurement from Minerva + first measurements of outgoing proton in T2K ND → both show clear discrepancy with available MC models  Moreover with Argon only protons/pions are accessible → measurement of neutrons is crucial for high energy (DIS) events and for all events with antineutrino (neutrino-antineutrino differences are at the core of d CP systematics!) 11

E n reconstruction: neutrons (2)  An example: ● smearing of E n CCQE is dominated by Fermi momentum Generator level ● smearing of E m +Ep is dominated by flux (and detector low binding E m +E p effects) energy → more robust estimator of E n against model biases high binding energy E n CCQE CCQE formula (i.e. constraining the model using the muon info only) But still important to correct for the right binding energy to get E n correct at the FD!  Fit to E m +Ep variable can extract the binding energy with very good precision at ND (depending on the precision of the flux → ~1 MeV ) different binding energy for proton and neutron → important to perform E m +En measurement at the ND to avoid n/n bias at the FD!  Need to be quantitative How well neutron measurements can be performed (at 3DST-like detector)? Spoiler: ~a factor two worse than proton (e.g. 2p2h with <5% precision) 12

Conclusions  With the huge stat + phenomenal amount of information on the final state of DUNE, we can move from model constrains to “data parametrization” of the model! the capability of measuring/disentangling IS/ s /FSI through their different A dependence is a crucial input to get the needed precision ● A-dependence of IS/ s /FSI can be driven by electron scattering data but need neutrino data at right energy (ND with Carbon + ND with Argon) to get the needed precision ● Carbon is an easier and well known nucleus → anchoring point to develop the constrains for Argon scattering the capability of measuring all the final state, including neutrons, is a crucial input to get the needed precision ● The hadron part of the final state is not enough well known to rely on the model for the n → n extrapolation Joint sensitivity studies are needed on a Carbon-target + Argon-target near detectors to be more quantitative 13